. 33
( 45 .)


W p7 passaged
22H l WA
79v aA n

301v A
22A l m

*a7v A
I t
100 200 300 400 500 600 700
lncubarion period (days)

Fig. 1. Incubation periods followrng lc inJection of 1% bram homogenates for 14
scraple/BSE strains in C57BL (Szncs7) (O), VM (5™incP7) (A), and C57BL x VM
(Sznc”˜P™) (M) mice *The incubation periods for C57BL and C57BL x VM mxe with
87V and for all three genotypes with 1llA are longer than 700 d and some mdlvlduals
do not develop chmcal disease wlthm their hfespan.

Thus, some agent strams produce shorter incubation periods in Sines7 mice
than m SincP7 mice, but this ranking IS reversed for other strains. Depending on
the agent stram, the mcubatlon period m the F1 heterozygote lies either between
or beyond those in the two parental genotypes. No agent strain has been iden-
tlfied with an incubation pertod in the heterozygote that IS shorter than in both
2.4. Neuropathological Differences: Lesion Profiles
Scrapie and BSE strams also can be distingmshed on the basis of the patho-
logical changes they produce m the brains of Infected mice. The most obvious
change seen m routme hlstologlcal sections is a vacuolation of the neuropil.
Agent strains show dramatic and reproducible differences in the severity and
dlstrlbutlon of this vacuolar degeneration m the brains of genetically uniform
mice. This forms the basis of a semiquantitative method of strain discnmina-
226 Bruce


Fig. 2. Standard brain levels and areas in which vacuolation is scored to produce
the lesion profile. Circled numbers indicate white matter areas.

tion in which the severity of pathology is scored from coded sections in nine
gray matter and three white matter brain areas (7) (see Fig. 2). Gray matter
vacuolation is scored on a scale of &5 and white matter vacuolation is scored
on a scale of O-3. The vacuolation score has been shown to bear an approxi-
mately logarithmic relationship to the number of vacuoles in a standard field
(7). For each group of mice the mean vacuolation score in each area is calcu-
lated to construct the “lesion profile.”
Each combination of agent strain and mouse genotype has a characteristic
lesion profile (I) (see Fig. 3). Unlike the incubation period, the lesion profile is
insensitive to the initial infecting dose and can be used to identify strains in
samples containing low levels of infectivity. This method can also be used
to determine which strain kills the mouse following injection with a mixture
of strains.
The distribution of vacuolar degeneration correlates closely with the distri-
bution of abnormal PrP accumulation in the neuropil, demonstrated by immu-
nostaining with PrP-specific antisera (8,9). These studies have demonstrated
clearly that pathological changes are targeted precisely to particular groups
Strain Typing Stud/es of Scrap/e and BSE

Scoring Area
Fig. 3. Examples of lesion profiles m VM mice for three strams of scrapie One to
mne are gray matter areas, 1*-3* are white matter areas

of neurons and that different strains of agent select different neuronal
3. Isolation of Strains of Spongiform Encephalopathies in Mice
3.7. Primary Transmissions to Mice
The strain-typing work at the Neuropathogenesls Unit has involved trans-
missions to mice from a wide range of natural and expertmental cases of
spongiform encephalopathy m sheep, goats, cattle, and a number of other spe-
cies. Because these diseasesare more difficult to transmit between species than
wlthm species (the “species barrier” effect) (IO), the inoculum is prepared from
a 10% rather than a 1% brain homogenate. Also, it has been found that mtra-
peritoneal inJection is more effective than intracerebral injection m some pn-
mary transmissions (II). In order to maximize the chances of a successful
transmission, mice therefore are injected by a combmation of routes: 20 pL
intracerebrally and 100 pL intrapentoneally. A larger panel of mouse strains IS
usually included m primary transmlsslon experiments than m subsequent
mouse-to-mouse passages because genes other than Sine may have a consld-
erable effect on incubation pertod at this stage (II) (see thefollowzng). In
general, the incubation periods in these experiments are very long and there
may be survivors. Some sources fall completely to transmit to mice of any

3.2. Serial Passage in Mice
Followmg a successful primary transmission, serial mouse-to-mouse
passage lines are set up routmely m C57BL (SzncS7) VM (Szncp7)mice,
using 1% bram homogenates from clmically affected indtviduals. At each pas-
sage the mcubation period and pathologrcal characteristics are tested m mice
of the three Szncgenotypes. On passage m a single inbred mouse strain, the
mcubation period usually shortens and stabilizes after a few passagesto give a
strain with characteristic properties. These properttes appear to be stable
mdefimtely on further mouse-to-mouse passage, as long as the conditions of
passage,particularly the mouse genotype, remain constant. In Edinburgh alone,
14 unequivocally distinct strains of scrapie and BSE have been isolated m
mice (see Fig 1). A further SIX isolates have unique disease charactertstics,
mdicattve of new strains, but are not yet fully stable; there is no indication
that we have reached the limit of this variation. It should be stressed that an
agent strain IS defined m terms of its characteristtcs, rather than its passage
history Thus, the same agent strain may be identified m a number of differ-
ent isolates.
Most primary sources give two different agent strains when passaged m
Sims7 and &ncP7 mice. These differences cannot be imposed by the host* A
number of different strains have been isolated in each mouse genotype, and the
same strain (e.g , ME7) sometimes has been isolated m both genotypes. How-
ever, the resolution of isolates into two distinct stable strains always is consis-
tent with the selection of strains that replicate more efficiently m the mouse
genotype used for passage. For example, 301C and 301V were isolated from
BSE, by passage in Sims7 and SzncP7 mice, respectively; 301C has a shorter
incubation period than 30 1V in Szncs7 mice and 30 1V has a shorter mcubation
period than 3OlC in Szncp7 mice (II) (see Fig 1).
3.3. Biological Cloning of Strains
Even when the disease characteristics have stabilized, it cannot be assumed
that the isolate contains only a single agent strain There is clear evidence that
minor strains can be passaged together with the major strain as a stable mixture
(2). For many experimental purposes this may not be important. However, for
some studies, particularly those concerned with the stability of strains when
passaged in different hosts, the presence of minor strains may become a seri-
ous complmatmg factor. To remove mmor strains, it 1snecessary brologrcally
to clone the isolate, by performing three or more sequential passagesusmg the
mmimum infecting dose. This IS achieved by mjectmg groups of mice with
serial tenfold dilutions of brain homogenate and passagmg from an affected
mdividual mjected with the highest dilution of moculum to produce clmical
cases.For most agent strains, thts hmitmg dilution is approx 10-˜. This proce-
Strain Typing Studies of Scrap/e and BSE
dure has been shown to lead to a permanent change m the charactertsttcs or
behavior of several isolates.
4. Stability of Mouse-Passaged Strains:
Detection of “Mutant” Strains
4.1. Stability on Passage in a Single Mouse Genotype
All but one of the scrapte and BSE strains listed m Fig. I appear to be stable
indefinitely when serially passagedat high dose m the mouse genotype m which
they were originally isolated. Seven of these strains (ME7, 22A, 22C, 22L,
79A, 87V, and 139A) have maintained then charactertsttcs throughout ten or
more serial passages in the appropriate mouse genotype. The one exception is
87A. The charactertsttcs of 87A are stable when it IS passaged at low dose m
Szncs7 mice, but often change suddenly when it IS passaged at higher dose m
the same mouse genotype, to give a strain with much shorter incubation peri-
ods (12). This new strain 1sstable, even when passaged at high dose, and always
IS Identical to ME7.
87A and ME7 produce very different patterns of pathology n-rthe brain and
it is posstble to classify each mdrvidual mouse on the basis of this pathology
In experiments using 1 or 10% 87A brain homogenates there ISa clear btmodal
distribution of incubation periods, with the ME7 type of pathology associated
with the shorter group (12). The changeover point is at approx 300 d, with little
or no overlap between the two groups. The proportion of mice m the shorter
group decreases with dtlution of the moculum, suggesting that the starting
homogenate contains a mixture of the two strains. Analysis of a large number
of experiments using different small brain samples to prepare the mocula has
shown that the ratio of ME7 to 87A is, on average, approx 1: 104.
The 87A isolate contains ME7 as a minor component, even after tt has been
cloned. This suggests that ME7 1s a shorter incubation period variant strain,
derived from 87A by a process analogous to mutation rn conventional microor-
ganisms. 87A has been isolated independently from stx different natural sheep
scrapte sources and m each case has behaved in the same way. This IS further
evidence that 87A and ME7 are closely related at the molecular level.
4.2. Stability on Changing the Mouse Genotype Used for Passage
From the preceding observations, tt is clear that scrapte-hke agents carry
some form of strain-specific information that IS independent of the host. Both
the “prton” and “vtrino” hypotheses for agent structure include the host com-
ponent, PrP. Therefore, an important question is whether the PrP genotype of
the host in which the isolate is passaged directly influences strain characteris-
ttcs. Thts question has been addressed by changing the Suzcor PrP genotype of
mouse used for serial passage, for several well-characterized scrapie strains
230 Bruce

Table 1
Summary of Experiments in Which Well-Characterized Scrapie Strains
Were Passaged Through Different Sine Genotypes of Mouse
Strams ongmally Isolated m Szncs7 mace
Sw2cp7 mice +
Uncloned 22C -+ Gradual change to 22H
Cloned 22C + Sm?™ mxe + Unchanged 22C
+ SwzcP™ mace + Unchanged ME7
Cloned ME7
Unchanged 139A
Uncloned 139A + SzncP™ mace +
Strains ortgmally isolated m SzncP™ mice
Gradual change to 22F
Uncloned 22A + Swzcs7 mace -+
Sm?™ mice +
Cloned 22A -+ Gradual change to 22F

have been tested m
At each passage tn the new host, the stratn charactertsttcs
the three Szncgenotypes Scrapte strains have been found to differ m the stabrl-
tty of their properties when passaged m this way (see Table I).
The charactertstics of some scraple strains (cloned ME7, cloned 22C, and
uncloned 139A) have been completely unchanged by passage m the alternate
mouse genotype. Other isolates (uncloned 22C, cloned or uncloned 22A) have
changed This type of experiment must be interpreted carefully to dtstmgutsh
between the three possible explanations for a change m properties
1 The selectton of a minor agent strain that was present already m the isolate,
2 The selectton of a vanant strain, derived by “mutatton” from the parental strain, and
3 A direct modtfcatton of strain determinants by the host

When such changes occur using uncloned isolates rt IS not posstble to deter-
mine then basis, unless other mformation IS available. For example, when
uncloned 22C IS passaged m SineP7mice its properties change gradually over
several passages to give a new strain, 22H (2). This can be interpreted only
with the knowledge that this change is not seen in the equivalent experiment
starting with cloned 22C (2). The concluston IS that 22C and 22H coexisted m
the early mouse passages of the isolate and that 22H was removed from the
mixture by cloning. These results emphasize the need to use cloned strains m
studies seeking host modtficattons of strain characterrstrcs.
On the other hand, when cloned 22A IS serially passaged m SznP™ mice, the
disease characteristics gradually change over several passages,eventually sta-
btlrztng to give the new strain, 22F (2). This suggests that 22F has been gener-
ated from 22A, either by a host-induced modrficatron or by a “mutatron” m the
mformatronal component of the agent. The fact that the change IS gradual
makes tt unlikely that the phenomenon 1ssimply owing to a modrficatton of the
agent by the host. The results are more consistent wrth the gradual selectron of
Strain Typing Studies of Scrapie and BSE
a mutant stram, 22F, which has a shorter incubation period m the new passag-
mg genotype than the parental stram, 22A. So far, there is no clear expertmen-
tal evidence that the Snc or PrP genotype of the mouse can actively modify the
charactertsttcs of any strain of agent.
4.3. Stability on Changing the Species Used for Passage
Strain characteristrcs can also be maintained on passage through another
species with a different PrP amino acid sequence. Ktmberlin and coworkers
have serially passaged a number of mouse scrapie strains in rats or hamsters
and then repassagedthe isolates in mice (13,14). They found that cloned 22A and
cloned ME7 were completely unchanged after serial passage m hamsters and
subsequent retsolatton m mice. Cloned 139A was unchanged by passage
through rats. In contrast, the propertres of cloned 139A and cloned 22C were
changed permanently by passage in hamsters, giving rise to new strains that
were stable on serial passage in mice. The latter results have been Interpreted
as the selection of strains in the new host spectes,other than the major strains
present m the ortgmal host. When this happens, it is likely to contribute to the
“species barrier” effect, i.e., the relatively long mcubatron period seen on pri-
mary transmissron to a new species, compared with the shorter mcubatton
periods seen on subsequent passage within this new species
5. Epidemiological Applications of Strain Typing
5.1. Strain Variation in Natural Scrapie
Because of the possibility of the selection of minor variant strains, it IS not
clear to what extent the mouse-passaged strains isolated from natural scrapre
cases are representative of field strains. However, strain typing m mice can
give some information about the extent of strain variation in the natural dis-
ease. In primary transmissions to mice, the successrate, incubation periods,
and pathology have varied enormously between sources. However, on further
serial passage in mice most UK isolates have given varying combinations of
the same three strains, 87A and ME7 m Szncs7 mice and 87V in Sncp7 mice.
Therefore, these sources may not be as diverse as they appear from the primary
transmission results. A series of transmissions from Icelandic sheep (15) have
given at least three strains in mice; these are not yet fully characterized but
clearly differ from strains from UK cases. In the United States, the mouse-


. 33
( 45 .)